Skip to content

Context-Efficient Reading Strategy

I need to... Guide Summary
Understand the core principle before picking a strategy Overview Use this guide before any code-reading task — before the first Read or Grep call — to pick the strategy that matches the task type instead of applying a blanket "save tokens" rule.
Fix a known bug or edit a known symbol Type A — Targeted Edits Use when you already know what symbol you're changing and the scope is bounded to one or two files. Bug fix, rename, small feature edit.
Document, audit, or review a module/package Type B — Comprehensive Understanding Use when the task requires understanding a whole unit — a module, package, library, or feature area. Documentation authoring, security/code review, architecture analysis, onboarding, multi-file refactoring.
Explore a large unfamiliar codebase Type C — Large Exploration Use when the task is open-ended, you don't know the structure yet, or a full Type B read would exceed ~30% of remaining context. Use Type B when the unit is bounded enough for main context.
Know what full reads catch that grep misses What Grep Misses Use when you're tempted to grep-first for a Type B task "to save tokens." This list explains what you'll miss and why that's fatal to documentation or audit accuracy.
Interpret a harness token-usage warning Token Warning Interpretation Use when the harness emits a warning like "Read results using 881.4k tokens (88%) → save ~264.4k" and you're unsure whether to change your reading strategy.
Verify a Type B task is actually complete Pre-Completion Checklist Use before declaring a Type B task complete — documentation finished, audit signed off, review approved. If any item was skipped because of context-saving instincts, the work is incomplete.
Look up the right strategy for a specific task quickly Quick Reference Use when you need the one-line answer for a specific task without reading the full guidance.